jeudi 4 juillet 2013

Training Intelligent Systems to Think on Their Own

Lehigh University (06/28/13) Kurt Pfitzer 

Lehigh University professor Hector Munoz-Avila has won a three-year U.S. National Science Foundation grant to develop autonomous agents that dynamically identify and select their own goals. Computing devices and software programs already use algorithms to make intelligent decisions, notes Munoz-Avila, who focuses on the incipient field of goal-driven autonomy (GDA) technology. In the near future, Munoz-Avila believes that algorithm-powered agents will analyze complex problems, establish the most effective intermediate goals, and take steps toward long-term solutions. During this process, agents will adjust to unexpected situations and learn from their mistakes without human intervention. "For a long time, scientists have told agents which goals to achieve," Munoz-Avila says. "What we want to do now is to develop agents that autonomously select their own goals and accomplish them." GDA agents could have useful applications in areas such as military planning, robotics, security, and electrical grid control systems. Munoz-Avila's project will focus on enabling GDA agents to acquire more knowledge of their domains and to generalize their success in other applications. He says the importance of GDA agents having the ability to diagnose environmental discrepancies and act intelligently is growing as autonomous computing devices and software proliferate in today's society.

Aucun commentaire:

Enregistrer un commentaire